Modified activated sludge processes have been used for the removal of nitrogen (N) and phosphorous (P), but in many cases the expected performance has not been obtained. Frequently, both N and P levels in treated sewage are found to be many times those encountered in similarly designed plants operating on different sources of sewage.
One well known design of nutrient removal process is the Bardenpho process, described in South African patent application 72/5371 to the South African Inventions Development Corporation (equivalent to U.S. Pat. No. 3,964,998).
For example, this process is known to operate well in one suburb of Johannesburg, but to perform very poorly in another. It has been postulated that sewage deficient in acetic acid, a well known fermentation product, is not suitable as influent to activated sludge plants designed for the removal of P. Therefore attempts have been made to improve the performance of plants by subjecting the influent raw sewage or the settleable portion thereof to fermentation with the objective of generating acetic acid (Barnard 1984).
This procedure has been found to be of assistance. However, the generation of acetic acid is limited by the well known fact that relatively slow growing bacteria are capable of degrading acetic acid to methane. Thus the system has previously been optimised to minimise loss of acetic acid by limiting the mean cell retention time in the pre-fermenter. Mean cell retention time (MCRT) is defined as the average time spent within the system by a cell. In the sewage treatment industry MCRT is usually called "sludge age".
Prior workers have recommended that mean cell retention times be restricted to 5 days or less in order to maximise the production of acetic acid (Nicholls et al. 1987, T'Seyen et al. 1985).